podcast-mcp

sgasser/podcast-mcp

3.2

If you are the rightful owner of podcast-mcp and would like to certify it and/or have it hosted online, please leave a comment on the right or send an email to dayong@mcphub.com.

Podcast MCP is a professional multi-speaker podcast generation tool integrated into Claude Desktop, utilizing high-quality AI voices and local processing for privacy.

Podcast MCP 🎙️

Generate professional multi-speaker podcasts directly inside Claude Desktop.

✨ Features

  • Real AI Voices: Uses high-quality voices (not robotic text-to-speech).
  • Multi-Speaker: Supports up to 4 distinct speakers automatically.
  • Zero Config: Works instantly after installation.
  • Privacy First: All processing happens locally on your machine.

🚀 Installation

Prerequisites

You need uv installed on your system via Homebrew (required for Claude Desktop to find it):

# macOS
brew install uv

# Linux/Other
# Use Homebrew on Linux, or ensure uv is in system PATH
# See: https://docs.astral.sh/uv/getting-started/installation/

Important: Installing via Homebrew ensures uv is in the system PATH that Claude Desktop can access.

Install the MCP Server

  1. Download podcast-mcp.mcpb from the releases.
  2. Drag and drop the file into Claude Desktop.
  3. That's it! 🎉

💡 Tips for Best Quality

  • Speed: If it feels too slow, increase TTS_SPEED in .env (e.g. to 1.3 or 1.5).
  • Pauses: You can adjust DEFAULT_PAUSE_MS in .env if the gaps between speakers are too long.
  • Remember: Restart Claude Desktop after changing .env for changes to take effect.

💡 How to Use

Simply ask Claude to create a podcast. You don't need to worry about formats or technical details.

Try this:

"Generate an audio podcast about the history of coffee."

Works in 20+ languages (German example):

"Erstelle eine Audio-Podcast-Datei über die Geschichte des Kaffees."

Claude will handle everything: writing the script, assigning voices, and generating the audio file.

⚙️ Configuration (Optional)

The default settings work great, but you can customize them if you want.

A .env file is automatically created on first run. To customize:

  1. Open the podcast-mcp folder (where Claude installed it).
  2. Edit the .env file to customize:
    • Voices: Choose from 50+ available speakers.
    • Speed: Adjust speaking rate (default is 1.2x).
    • Output Folder: Change where files are saved (default: ~/Downloads).
    • Log Level: Set to DEBUG for detailed logs (default: INFO).
  3. Restart Claude Desktop for changes to take effect.

❓ FAQ

Where are my podcasts saved? By default, they appear in your Downloads folder.

Can I use my own voices? Currently, we support the built-in high-quality voices from Coqui XTTS-v2.

Does it work offline? Yes! The first time you run it, it downloads the AI model (~2GB). After that, it works entirely offline.

Where can I find debug logs? Claude Desktop automatically captures all logs to ~/Library/Logs/Claude/mcp-server-Podcast MCP.log. You can view them with:

tail -f ~/Library/Logs/Claude/mcp-server-Podcast\ MCP.log

To change the log level, edit MCP_LOG_LEVEL in .env (options: DEBUG, INFO, WARNING, ERROR).

🙏 Acknowledgments

This project is built on excellent open source software:

  • Coqui TTS - High-quality text-to-speech with XTTS-v2 model
  • MCP - Model Context Protocol for Claude integration
  • pydub - Audio processing and manipulation
  • uv - Fast Python package manager
  • python-dotenv - Environment configuration management

License: MIT | Author: Stefan Gasser